What is the most important function of the adolescent peer group?

aDepartments of Preventive Medicine and Psychology, Institute for Health Promotion and Disease Prevention Research, University of Southern California, 1000 S Fremont Avenue, Alhambra, CA 91803, USA

Find articles by Steve Sussman

Pallav Pokhrel

aDepartments of Preventive Medicine and Psychology, Institute for Health Promotion and Disease Prevention Research, University of Southern California, 1000 S Fremont Avenue, Alhambra, CA 91803, USA

Find articles by Pallav Pokhrel

Richard D. Ashmore

bDepartment of Psychology, Rutgers, The State University of New Jersey, 152 Frelinghuysen Road, Piscataway, NJ 08854, USA

Find articles by Richard D. Ashmore

B. Bradford Brown

cDepartment of Educational Psychology, University of Wisconsin-Madison, 1025 W Johnson St., Madison, WI 53706-1796, USA

Find articles by B. Bradford Brown

Author information Copyright and License information Disclaimer

aDepartments of Preventive Medicine and Psychology, Institute for Health Promotion and Disease Prevention Research, University of Southern California, 1000 S Fremont Avenue, Alhambra, CA 91803, USA

bDepartment of Psychology, Rutgers, The State University of New Jersey, 152 Frelinghuysen Road, Piscataway, NJ 08854, USA

cDepartment of Educational Psychology, University of Wisconsin-Madison, 1025 W Johnson St., Madison, WI 53706-1796, USA

*Corresponding author. IPR-USC, 1000 South Fremont Avenue, Unit 8, Alhambra, CA 91803, USA. Tel./fax: +1 626 457 4012. ude.csu@amssuss (S. Sussman)

Copyright notice

The publisher's final edited version of this article is available at Addict Behav

Abstract

This study provides an exhaustive review of 44 peer-reviewed quantitative or qualitative data-based peer-reviewed studies completed on adolescent peer group identification. Adolescent peer group identification is one’s self-perceived or other-perceived membership in discrete teenage peer groups. The studies reviewed suggest that adolescent peer groups consist of five general categories differentiable by lifestyle characteristics: Elites, Athletes, Academics, Deviants, and Others. We found that the Deviant adolescent group category reported relatively greater participation in drug use and other problem behaviors across studies, whereas Academics and Athletes exhibited the least participation in these problem behaviors. Additional research is needed in this arena to better understand the operation of adolescent group labels.

Keywords: Adolescent group identification, Group characteristics

1. Introduction

Social scientists have long noted the tendency for people to place themselves and others into consensually recognized and labeled social types (Ashmore, Del Boca, & Beebe, 2002). Adolescentstend to segregate themselves into different peer group types. Adolescents give names to their peer group types, as has been popularly illustrated by movies such as The Breakfast Club (1985) and Clueless (1995). Peer group names that adolescents give themselves or each other suggest the groups’ lifestyle characteristics, such as shared beliefs, interests in clothes and music, and preference for specific activities (Brown & Lohr, 1987; Hartup, 1985; Sussman et al., 1990). As discussed by Brown and Lohr (1987), adolescents may identify with groups to develop a sense of identity and a positive self-concept, and an increased sense of personal autonomy from parents. In addition, these group categories may reinforce cultural norms by indicating successful and unsuccessful ways of participating in the culture (Ashmore et al., 2002).

Teens may “place” themselves into peer group types in at least two ways. First, they may simply identify themselves with a certain peer social type regardless of any direct interaction with other peers. In this sense, they are making a statement about the type of teen they are within the culture (i.e., they are stating the name of the reputation-based collective in which they feel they take part). Second, these adolescents may actually participate in peer groups which reflect the larger collective. The peer groups provide a check on whether they view youth as “really” a member of the peer group type or only someone who tries to be part of the group (a “wannabe”). Adolescents are in the process of moving away from the closed environment of the parental home where they are largely influenced by their immediate family to a social world where they are among peers and have to begin to make independent choices. Due to lack of experience they are often not sure about the lifestyle decisions they should make (e.g., balancing their social and school lives, vocational orientation). In need of support and direction they are likely to search for a place among a group of peers by conforming to the group’s norms (Larkin, 1979). Peer groups thus either vicariously or directly facilitate the adolescents’ transition into the larger social environmental world.

The literature that pertains to the study of adolescent peer group types has been referred to by various names (e.g., peer group association (Sussman et al., 1990), peer group self-identification (Sussman et al., 1994), peer crowd affiliation (Prinstein & La Greca, 2002)). For the remainder of this review, we will label this arena as “peer group identification” (e.g., Mosbach & Leventhal, 1988) because peer group types may be self- or other defined, and may pertain to a larger collective or to actual peer group interactions. This term permits inclusion of that variation. The peer group identification literature began at least four decades ago (Clark, 1962). Several studies have found that peer group identification is related to problem-prone behaviors such as substance use and risk-taking (see Jessor, 1984). Findings across several peer group identification studies also suggest that a social hierarchy exists among adolescent groups with Elites or Athletes at the top, and that this hierarchy is associated with the level of one’s social involvement, social acceptance, or self-esteem (e.g., Brown & Lohr, 1987; Cohen, 1979; Eder, 1985; Franzoi, Davis, & Vasquez-Suson, 1994; La Greca, Prinstein, & Fetter, 2001; Prinstein & La Greca, 2002).

Currently, it is unknown (1) how many data-based peer-reviewed studies on youth peer group identification exist, (2) what the variation of methods are that delineate peer group names, and (3) what the patterns of associations are between peer group identification and behavioral (e.g., drug use) or personality (e.g., self-esteem) variables across studies. The purpose of this review paper is to address these questions. We first attempted to identify all quantitative or qualitative data-based peer-reviewed studies that used peer group names to identify peer groups. Then, we examined the methods used to delineate group names. We also attempted to identify a finite number of general group names that accommodate the studies that were completed. Two recent group type scaling studies (Ashmore, Griffo, Green, & Moreno, in press; Stone & Brown, 1999) provided direction on what the general group labels might be. We used the results of these two studies and used a sorting procedure, both of which are described below, to identify general groups.

Next, the relations of peer group identification with general group characteristics, personality characteristics, parenting characteristics, substance use (tobacco, alcohol, marijuana and illicit drugs), teen dating/sexual behavior, and violence participation were examined. It was expected that youth who were identified as belonging to a Deviant group (e.g., a Stoner, Tough, Heavy Metaler, Druggie, or Burnout, as examples), would be delineated across studies and would show the greatest prevalence of problem behavior and personality characteristics (Jessor, 1984). On the other hand, we expected that youth that were more involved in school activities or social events, such as the Academics, Elites, or Athletes would show the lowest prevalence of drug use behavior among the groups. Also, we expected that the Others, not falling into a clearly defined group, perhaps lacking peer support to boost their sense of self-worth, would also show higher drug use than Academics, Elites, or Athletes. Finally, limitations of this research and future research directions based on this review were suggested to stimulate continued investigation in the field of group identification.

2. Method

2.1. Selection of studies

Searches of OVID MedINFO (1966 to August, 2005), PsycINFO (1887 to August, 2005), ERIC (1966 to August, 2005), Social Science Abstracts (1983 to August, 2005) and Sociological Abstracts (1963 to August, 2005) databases were completed to identify studies that used peer group identification to delineate adolescent peer groups. All databases were searched by crossing the terms “youth” and “adolescents” with “peer group identification”, “peer group self-identification”, “peer group association”, “peer group”, “peer crowd”, “peer crowd affiliation”, “peer affiliation”, and “peer group affiliation.” In addition, we engaged in searches using the group names “Jock”, “Brains”, “Elites”, “Socies”, “Druggies”, and “Normals.” Only studies published in English language peer-reviewed journals were included. Studies were also excluded from consideration if they concerned subjects more than 26 years of age (end of emerging adulthood (Arnett, 2000)) and if they did not identify clearly definable peer group names. Forty-four studies were located. The included studies and the group names used in them are shown in Table 1.

Table 1

Group identification studies and group names

Peer group names

Open in a separate window

2.2. Subjects

Subjects across these studies varied in age from 9 to 23 years (see Table 2). While 30 studies involved solely general populations of junior or senior high school students (grades 7 to 12; 12–18 years of age), 3 studies (Sussman et al., 1999; Sussman, Dent, & McCuller, 2000; Sussman, Unger, & Dent, 2004) involved alternative (continuation) high school students (for those that do not successfully adapt to mainstream education). Also, 4 studies targeted pre-teens, at least in part (Dubow & Cappas, 1988; Eder, 1985; Michell, 1997; Prinstein & La Greca, 2002). One study (Ashmore et al., 2002) addressed college students. Six studies (Downs & Rose, 1991; Kipke, Montgomery, Simon, Unger, & Johnson, 1997; Kipke, Unger, O’Connor, Palmer, & LaFrance, 1997; Miller, Farrell, Barnes, Melnick, & Sabo, 2005; Miller et al., 2003; Sussman et al., 1999) targeted non-school based samples. Only three studies were conducted outside the U.S., all in school-based settings (Heaven, Ciarrochi, Vialle, & Cechavicuite, 2005; Michell, 1997; Thurlow, 2001).

Table 2

Peer group differentiation according to methods and measures

StudyMethodResultsSubjectsAshmore et al. (2002)Other-peer ratings of groups (only perceptions of group types)Group types devoted to sexual and social pursuits (e.g., Frat boys) were more likely to use alcohol than those more involved with academic pursuits (e.g., Brains)300 university students (mean age=19.2) from 1 school, NE, 56% F, 45% W, 10% B, 10% H, 30% A, 5% OBarber et al. (2001)Self-identificationJocks and Criminals reported more drinking than other groups; Criminals used marijuana most often; Brains were most likely to have graduated by 24 years old; Basket cases and Criminals reported the lowest self-esteem900 Michigan Study of Adolescent Life Transitions participants (assessed at 10th, 12th grades and 2, and 6 years after high school), from 10 school districts, MW, F NR, mostly WBrown and Lohr (1987)Other-peer ratings of group types, friends in groups, and self-identificationEnvious and Distorters showed significantly less self-esteem when compared to Jocks and Populars327 7th–12th graders from 1 JHS and 1 RHS, MW, 48% F, 98% WBrown et al. (1993)Other-peers selected by school staff; Other-peer ratings of group types, and peers into groupsGood parenting practices, academic achievement, and self-reliance positively associated with Jocks, Populars, and Brains and negatively associated with Druggies; Druggies positively related with controlled substance use3781 9th–12th graders, 3 MW and 3 We RHS, 52% F, 61% W, 12% B, 13%H, 12% AClasen and Brown (1985)Other-peers selected by school staff; Other-peer ratings of group types, and peers into groupsDruggies/Toughs greater peer pressure towards misconduct (e.g., drug and alcohol use); Jocks greater peer pressure towards school involvement; Loners greater peer pressure towards school involvement in rural areas but lower than Jocks in urban areas689 7th–12th graders from 2 middle schools and 2 RHS, 2 MW communities, half rural and half urban, 50% F, 95% WCohen (1979)Self-identificationFun Group athletic, engaged in extracurricular activities, popular; Academic group academically oriented; Delinquent group rejects studies and dates, smokes and drinks; Nobodies as “almost faceless student who never speaks up”1038 9th–12th graders from 1 RHS, MW, 51% F, 100% WDemuth (2004)Self-identificationLoners less delinquent than Non-loners; Loners less likely to go on dates, to parties, or to other social activities than Non-loners; Loners’ friends more approving of delinquency than Non-loners’ friends1237 13- to 19-year-old participants of National Youth Survey, 46% F, 79% WDolcini and Adler (1994)Other-peers selected by school staff; Other-peer ratings of group types, and peers into groupsElites more athletic than Smarts or Outsiders; Smarts more academically competent than Outsiders; Popular Blacks (mostly female) resembled Elites and showed higher perceived social competence than Smarts, Floaters, and Outsiders; no differences among groups on physical appearance, behavioral conduct, or global self-worth; Elites and then Popular Blacks most likely to smoke (cigarettes), drink, use marijuana, and report sexual intercourse183 8th graders from 1 middle school, We, 52% F, 20% W, 24% B, 19% H, 20% A, 17% ODowns and Rose (1991)Self-identificationBrains/Smart Ones intellectual and adverse to alcohol and drug use; Jocks/Socies value self- image as reflected in personal appearance, clothes, popularity, athletic skills, good grades, and going to college; Average/Normals moderately involved in school activities, more inclined to use alcohol and drugs than Brains and Jocks; Druggies/Losers/Heads/Rejects least involved in school activities and low position in the status hierarchy; and highest use of alcohol and drugsTreatment sample: 127 13- to 17-year olds from 1 hospital-based program, MW, 49.4% F, 96% W; Comparison sample: 114 13- to 17-year olds telephone-selected, MW, 46% F, 99% W (samples combined for analysis)Dubow and Cappas (1987)Teacher ratings of students; Other-peer ratings of peers into groupsTeacher ratings: Rejected fewest friends, lowest GPAs, and most problem behavior; Popular and Controversial best adjusted to school; Peer reports: Rejected and Neglected least competent, Popular and Controversial most competent; Rejected and Controversial behavior problems compared to Popular and Neglected; Self-ratings of friends and self-esteem: Rejected reported fewest friends but no difference in self-esteem238 3rd–5th graders from 4 elementary schools, MW; 140 students urban, 50% F, 51% W, 49% B; 98 students rural, 60% F, 99% W, 1% BDurbin et al. (1993)Other-peers selected by school staff; Other-peer ratings of group types, and peers into groups, and self-identificationAdolescents with highest GPAs and from authoritative parenting style families in well- rounded peer groups (i.e., the Jock, Brain, Popular, and Average); Nonsocials (i.e., Loners and Nerds) also highest GPAs; boys from “indulgent” families oriented toward Partier group; girls from uninvolved families overrepresented in the Druggies and Partier groups; Druggies and Partiers most likely to use drugs3407 9th–12th graders from 9 RHS, MW and We, 53% F, 100% WEckert (1983)Ethnography (interviews and participant observation)Jocks more positive roles at school, more likely to be from higher SES background, and less likely to smoke (cigarettes); Burnouts less likely to participate at school, more likely to smoke, and more likely to be from lower SES background200 high school students from 1 RHS, MW, 50% F, mostly WEder (1985)Ethnography (interviews and observation)Elite group most visible; cheerleaders became popular at first but later became increasingly disliked for being snobs and stuck-up (cycle of popularity)750 6th, 7th, and 8th graders from 1 middle school, MW; mostly F and W in groups observedEicher et al. (1991)Ethnography (interviews and participant observation)Jocks wore letter jackets, jerseys, nice jeans, gym shoes; Nerds wore out-of-style clothes, unkempt hair; Punks and Freaks partly shaved, spiked-up hair, wore black leather items; Preppies wore nice, expensive clothes; Average/Regulars dressed casually10th graders from 1 RHS (11 students interviewed), MW, F and ethnicity NRFarmer et al. (2002)School staff-rated, and youth generated social networksAggressive groups, and Isolates, most likely to dropout from school; Popular groups (among males only), and Zero-popular groups also relatively likely to dropout475 7th graders, 3 SE communities, 52% F; 70% W, % O NRFishkin et al. (1993)Perception of group types, and self- identificationHigh risk youth were perceived by their peers (contrary to their own group perception) to be less engaged in low risk school and non-school activities, more engaged in drug use, and less likely to find good jobs later340 7th graders from 3 JHS and 615 10th graders from 2 RHS, We, 49% F, 58% W, 27% H, 5% B, 5% A, 15% OFordham and Ogbu (1986)Ethnography (interviews)Underachievers avoid being accused of “acting white”, to be a Kin/Regular Black; successful students use methods to cope with being accused of “acting white”; “Brainiacs” are those that sell out to white society; “Pervert- brainiacs” totally sell out and questionable in their sexuality33 11th graders from 1 RHS, NE, 50% F, 99% B at schoolFranzoi et al. (1994)Other-peer ratings (up to 10 other students most and least desire to spend time with)Popular and Controversial most attractive, highest grades, greatest social participation and athletic achievements; Average less attractive and less social participation, fewer social honors; Neglected/Rejected, compared to Average, dated less, least popular and athletic, most lonely408 9th–12th graders assessed in 2 consecutive years at 1 RHS, MW, 49% F, mostly WGotlieb (1975)Other-peer rating (14 raters)Elites worry about school performance, plans after graduation, fear of disappointing parents; Isolates worry about relations with peers, family members, girls, and lack of motivation/ passivity; Deviants perceive needing to understand themselves better and deal with their drug use; Outsiders most likely to work and worry about work20 12th graders (5 from each group) from 1 RHS, MW, ethnicity NRHeaven et al. (2005)Self-identificationRebels most likely to provide depressive explanations for events, least likely to experience authoritative (democratic) parenting; Studious group on other extreme, but not different from groups other than rebels893 12-year olds from 6 Catholic HS in New South Wales, Australia, 49% F, %W NRKinney (1993)Ethnography (interviews and observation)Nerds (unpopular, studious, low social skill) felt like “outsiders” or “social outcasts” back in middle school, avoided by Popular groups (e.g., Trendies); Nerds experienced transition to high school positively (gained self-esteem), opportunity for membership in greater variety of groups, less pressure to appear popular, and popular groups look down at other groups too (e.g., punk rockers)81 9–12th graders from 1 RHS, MW, F NR (both M and F interviewed), mostly WKipke, Montgomery, et al. (1997)Self-identificationPunks and Gay/Bisexual groups most likely to report substance abuse disorders (Gay/Bisexual for alcohol, not other drugs); Punks most likely to have unprotected sex and share needles; Gay/ Bisexuals and hustlers most likely to engage in survival sex; Hustlers most likely to be tested for HIV; Students/Athletes lowest risk for an alcohol use disorder303 13- to 23-year olds (73% 18 or older), We, homeless or at imminent risk of homelessness, 34% F, 54% W, 14% H, 19% B, 13% OKipke, Unger, et al. (1997)Self-identificationPunks/Skinheads and Druggies out on the street longest time; Punks/Skinheads most likely to have used services at drop in centers and rely on panhandling for subsistence; Hustlers most likely to rely on prostitution for subsistence; Gang members most likely to use shelter services, rely on drug dealing or stealing for subsistence, and on the street for shortest time752 12- to 23-year olds, street youths (50% 19 or older), We, 29% F, 52% W, 15% H, 19% B, 2% A, 11% OLa Greca and Harrison (2005)Self-identificationHigh status groups (Populars, Jocks) less social anxiety and depression than Others; affiliation with low-status groups (Burnouts, Alternatives) not related to social anxiety or depression; any group identification less anxious421 10th–12th graders from 1 RHS, SE, 57% F, 17% W, 67% H, 9% B, 6% OLa Greca et al. (2001)Self-identificationBurnouts and Nonconformists, and friends, greatest drug (smoking, alcohol, marijuana, other drugs) and # sex partners, relatively low social acceptance; Brains, least health-risk activities (low social acceptance); Jocks and Populars, and friends, high social acceptance, lower substance use, but Populars higher on alcohol use, Jocks marginally higher on casual sex250 10th–12th graders that had been tracked from 3 elementary schools, SE, 60% F, 46% W, 37% H, 13% B, 4% AMatute-Bianchi (1986)Ethnography (interviews and observation)Successful Japanese-descent students more knowledge about adult opportunities and relation of school to post-secondary experiences than successful Mexican-descent students; Chicanos and Cholos more gang-oriented, and more likely to be enrolled in alternative programs designed for unsuccessful students35 Mexican-descent and 14 Japanese- descent students from 1 RHS (school: %F NR, 33% W, 57% H, 9% A, 1% B), WeMichell (1997)Ethnography (focus groups and interviews)Top boys/girls popular, in fashion; Top girls smoked, drank, used hashish more than Top boys, liked boys a lot; Top boys liked sports; Middle group studious and future-oriented, not smoke or peer pressure to smoke; Bottom low- status pupils bad grades, truancy, peer pressure to smoke; Bottom trouble-makers hated school, into fights, smoking, drinking, marijuana use, other drug use; Loners not smoke36 11-year olds, and 40 13-year olds, primary and secondary school in West Scotland, %F and ethnicity NRMiller et al. (2003)Self-identificationMale and female Jocks scored in sports participation but also higher than Non-jocks on level of drinking; Male Jocks reported higher frequency of alcohol-related social problems compared to male Non-jocks699 households, 13- to 16-year olds, NE, 48% F; 70% W/O, 30% BMiller et al. (2005)Self-identificationMale Jocks frequent dating, earlier sexual debut, more frequent past-year and lifetime sexual activity, higher lifetime total # sex partners compared to Non-jocks; Jocks greater past year hours athletic activity600 households, 12- to 17-year olds, NE, 54% F; mostly B and WMosbach and Leventhal (1988)Self-identificationDirts highest in risk taking preference, most likely to smoke, drink coffee, drink alcohol and date; Hot-shots more likely to smoke and drink hard liquor than Jocks and Regulars; Regulars highest in self-esteem353 7th–8th graders from 1 JHS, MW, 46% F, ethnicity NRPascoe (2003)Ethnography (interviews and observation)Jocks at highest ranked position in social order; male Jocks distinct (female Jocks dispersed through multiple groups); Jocks associated with dominant masculinity; athleticism is treated as “insurance” for masculinity and Jock membership; those who reject the Jock dominant hierarchy exhibit “counterculture” identity (e.g., Freaks).20 15- to 18-year olds from 2 RHS, We, 0% F for the study, ethnicity NRPoveda and Crim (1975)Other-peers selected by school staff; Other-peer ratings of group types, and peers into groupsHigh society girls more likely to participate in school activities than Party girls; High society girls further divided into Pep crews and Snobs; Pep crews represent ‘athletic spirit’; Snobs are fashion setters and less actively involved in school compared to Pep crews; Party girls likely to smoke, drink, cruise, engage in sex and use drugs; Extreme Party girls called ‘Dopers’400 12th graders from 1 RHS, We, 100% F for the study, W majority.Prinstein and La Greca (2002)Self-identificationPopulars/Jocks highest physical appearance, social acceptance, athletic ability, romantic appeal, global self-worth, and least depressed affect, social anxiety, and loneliness; Brains highest scholastic competence, but only Brains showed a decrease in self-esteem and increase in loneliness over time (not extreme rating at either time point, as Deviants were the lowest in self-esteem at both time-points); None/average more romantic appeal than Brains; Burnouts lowest levels of competence in behavioral conduct, most depressed affect246 10th–12th graders (across 25 RHS, as a 6-year follow-up cohort from 3 elementary schools), SE, 60% F, 46 W, 37% H, 13% B, 5%ARiester and Zucker (1968)Self-identificationIdentification with either Collegiates or Leathers associated with high drinking; parent use of alcohol significantly associated with teen drinking; drinking reported more in group setting than while alone; Intellectuals highest grade and least likely to drink143 11th and 12th graders from 1 high school, NE, 50% F, ethnicity NRSussman et al. (1990)Self-identificationDirts and then Skaters most likely to smoke cigarettes; Dirts and Skaters most likely to use smokeless tobacco, and lower grades than Hot- shots and Jocks; Dirts highest in risk taking, most likely to try alcohol, marijuana, and hard drugs, least involved in sports, lowest in self- esteem; Hot-shots least likely to smoke; Jocks most involved in team sports340 7th graders from 3 JHS and 615 10th graders from 2 RHS, We, 49% F, 58% W, 42% mostly HSussman et al. (1993)Self-identificationHigh risk youth most likely to be weekly cigarette smokers, sibling and best friend smokers, stick to group even if meant trouble, lie to protect friends, get revenge or party as coping strategies; High risk youth least likely to be in sports, feel that school reputation is important, though not differ in self-esteem from other groups; non-smoking High risk youth unlikely to have close friend smoker and likely to place importance on health as value1245 9th–12th graders from 12 RHS, We, 48% F, 59% W, 21% H, 20% OSussman et al. (1994)Self-identificationGroup self-identification (High risk youth group status) in 7th grade predicted smoking in 8th grade (but not the converse); compared with 7 other psychosocial predictors, group self-identification as good a predictor (not a mere proxy); High risk youth highest group stability over time.3750 7th graders from 20 JHS, We, 50% F, 60% W, 27% H, 7% B, 6% A/OSussman et al. (1999)Self-identificationGroup self-identification significant predictor of cigarette smoking, alcohol use, marijuana use, and # types illegal drugs used across samples; comparable to demographics and 6 psychosocial variables, not a mere proxy of other variables across samples; High risk youth more likely to use on drug measures compared to Jocks, Regulars, Others, and Hot-shots; street youth 3 groups with Regulars/Jocks/Hot-shots merged3061 9th graders from 34 RHS, 803 14- to 19-year olds from 21 CHS, 425 13- to 23- year old street youth; We; 50% 38%, and 34% F; 49%, 39%, and 51% W; 29%, 41%, and 15% H; 22%, 20%, and 34% B/A/OSussman et al. (2000)Self-identificationGroup self-identification 1-year prospective predictor of violence perpetration, victimization, fear of victimization, weapons carrying, and cigarette, alcohol, marijuana, hard drug use, drug use intent, drug abuse, and hangout and anger coping; High risk youth highest on these measures compared to Hot-shots, Others, and Regulars681 14- to 19-year olds from 21 CHS, We, 45%F, 36% W, 48% H, 8% B, 8% A/OSussman, Unger and Dent (2003)Self-identificationFive years later, baseline-identified High risk youth most likely to use hard drugs, suffer drug use consequences, perpetrate violence, become victimized, and least likely to graduate from high school and find stable employment (compared to baseline Regulars, Hot-shots, and Others); Others and High risk youth most likely to use alcohol and marijuana, receive financial aid, and be involved in drug-related driving; Regulars and High risk youth most likely to be parents532 19- to 24-year olds previously from 21 CHS, We, 43% F, 31% W, 50% H, 6% B, 5% A, 9% OThurlow (2001)Self-identificationMost peer groups similar to those in U.S. (e.g., Populars, Brains, Trouble-makers, Hard people/Toughs, Others/normals) whereas two notable groups, Jocks and Burnouts, not self- identified in this British sample (4% were in sports groups)462 14-year olds from 6 high schools in England and Wales, 48% F, 74% W, 26% OTolone and Tieman (1990)Statistical method (1st and 4th quartiles of self-reported Social Involvement Index used to identify groups)Socials more likely to engage in alcohol and drug use, truancy, delinquency and violence, but also more involved in creative writing and more happy and satisfied with life, than Loners; Loners more likely to engage in conventional activities such as reading and listening to music10,862 12th graders from randomly selected schools across the US., 46% F, 100% WUrberg (1992)Self-identificationBurnouts likelytosmoke weekly most, followed by Averages and Jocks/Preps; Burnouts lower in conformity to peers than either Jock/Prep or the Average groups; best friends major source of influence on smoking compared to group324 11 graders, MW, 51% F, 96% WUrberg et al. (2000)Other-peer ratings of students into groups and group self-identificationConcordance (overall 66%) between self- identified and peer-identified groups was greater for Alternatives (100%) and Burnouts (64%) than Jocks (56%), Preps (53%), Average (53%), Whiggers (33%), Brains (14%), and Nerds (2%; n=5); Brains highest GPAs, lowest delinquency and drug use (self- and other- rated); Whiggers, Burnouts, and Alternatives most delinquency, cigarette, alcohol, marijuana use and lowest grades (self- and other-rated); Preps, Jocks, and Average between extremes on all measures; Nerds like Brains on drug use but grades like Preps, Jocks, and Average489 7th, 9th, and 11th graders, MW, 50% F, 92% W

Open in a separate window

Notes: JHS=junior high school; RHS=regular high school; CHS=continuation high school; MW=Midwest; SE=Southeast; NE=Northeast; We=West; F=females; W=White; B=Blacks/African Americans; H=Hispanic; A=Asian/Pacific Islanders; O=Other; NR=not reported.

2.3. Assessing peer group identification and characteristics

Table 2 lists the methods used to identify peer group names and characteristics. The methods used by the majority of studies to elicit peer group names and their respective characteristics are differentiated in three basic ways: (1) adolescents’ self-report on their own peer group identification (24 studies), (2) investigators’ classifying of adolescents into peer groups based on use of ethnographic methods — unstructured interviews and participant or naturalistic observation (eight studies), and (3) peer ratings of adolescents into groups according to the perceived ‘social types’ prevalent at their schools (seven studies).

Five other studies were rated differently. One study only investigated perceptions of group types, not dividing subjects into groups (Ashmore et al., 2002). Four studies used statistical methods to divide other-report or self-report measures of aggression, social network ratings of peers one desires to spend time with, or social involvement to create groups (Dubow & Cappas, 1988; Farmer et al., 2003; Franzoi et al., 1994; Tolone & Tieman, 1990).

2.4. Creating general group name categories across studies

After locating all peer-reviewed data-based group identification studies, we organized them by placing their groups into a general group name framework. To create this framework we relied on two sources of information. First, we examined the results of two scaling studies that placed perceptions of group types into two-dimensional space. A social map was created by Stone and Brown (1999) derived from a multidimensional scaling of perceptions of general group types on peer status versus academic engagement, among a sample of approximately 2000 students grades 6–12 who did the ratings. Ashmore et al. (in press) used cluster analysis to scaled perceptions of specific college student group types on social involvement versus academic involvement. The specific groups clustered to reveal general categories that aligned at different locations on these dimensions. Their results replicated the ones obtained by Stone and Brown (1999).

Based on this previous group name perception work, we conjectured that youth could be grouped into one of 5 general categories: Elites, Athletes, Deviants, Academics, and Others. Elites are high in peer status and social involvement, and somewhat involved in academics. Athletes are high in peer status and social involvement but only slightly involved in academics. Deviants are in the middle on peer status and social involvement and rebel against school (very low academic involvement). Academics are high in academic involvement, in the middle on peer status, and relatively low on social involvement. Finally, Others tend to be relatively low in peer status, social involvement, and academic involvement.

Second, we engaged in a rating task. The specific group names used to compose general group names were extracted from each study. These names were placed on index cards. If two names were nearly identical (e.g., “Socies”, “Socials”, “High society”) these names were placed on the same card. A total of 119 cards were created. Next, 4 raters (2 males and 2 females, 22 to 30 years old) independently were requested to place the 119 cards into the 5 categories created by the authors. The 5 general categories were defined by use of one sentence descriptions, each. Percent agreement of the placement of the cards into groups by each rater with the arrangement of the names in Table 1 (the standard, created by the authors) was calculated. Percent agreement varied from 81% to 90% (mean across raters=85.0%).

3. Results

3.1. Group names used in the study dataset and their general lifestyle characteristics

3.1.1. The Elites

The Elites category was recognized across studies as being a high status group (i.e., in the 34 studies that delineated this category). An elite-type group was the leading group at school and comprised of members who generally were successful in academic and extracurricular activities, held a high opinion of themselves, and were high in both other-perceived and self-perceived social competence. The 10 studies that failed to delineate Elites included seven which subsumed Elites under an Athlete label (Downs & Rose, 1991; Eckert, 1983; Fordham & Ogbu, 1986; Kipke, Montgomery et al., 1997; Kipke, Unger et al., 1997; Riester & Zucker, 1968; Urberg, 1992), and three studies that did not attempt to study Elites per se (they compared Athletes to Non-athletes, or Loners to Non-loners (Demuth, 2004; Miller et al., 2003, 2005)).

3.1.2. The Athletes

Athlete group status was associated with being popular (as well as athletic) like the Elite category (in all 33 studies that this group was mentioned); in fact, seven studies used an Athlete group name to refer to the popular group at school, as was just described in the Elites subsection. Athletes were not represented in 11 studies. Among these, five studies referred to Elite youth as also being the athletes in the school (Cohen, 1979; Dolcini & Adler, 1994; Gotlieb, 1975; La Greca & Harrison, 2005; Michell, 1997). (Thus, 12 studies either subsumed Elites within the Athletes category, or subsumed Athletes within the Elites category.) Also, five studies focused on comparisons that would tend to subsume Athletes as part of a larger group (Loners versus Non-loners (Demuth, 2004; Tolone & Tieman, 1990); Rejected versus Non-rejected (Dubow & Cappas, 1988; Franzoi et al., 1994); Aggressive versus Non-aggressive (Farmer et al., 2003)). Finally, in one study of Mexican and Asian immigrants (Matute-Bianchi, 1986) no Athletes were described at the study high school.

3.1.3. The Academics

Clark (1962) described “Academics” as a peer group more devoted to academic studies and academic-related extracurricular activities than to any other activities. They have been found to be most likely to have graduated from college by 24 years old (Barber, Eccles, & Stone, 2001). A total of 17 studies failed to delineate academics among the groups. Of these studies, four subsumed Academics under the Elites category (Mosbach & Leventhal, 1988; Pascoe, 2003; Poveda & Crim, 1975; Sussman et al., 1999), one study subsumed Academics under the Others category (La Greca & Harrison, 2005), one only delineated two groups in the school and Academics and Elites were subsumed under an athletes category (Eckert, 1983), five focused on types of a priori group comparisons which would tend to group Academics together with other groups (e.g., Demuth, 2004; Miller et al., 2003, 2005; Sussman et al., 1993; Tolone & Tieman, 1990), and six focused on single characteristics of groups and failed to identify Academics as a separate group (clothing styles (Eicher, Baizerman & Michelman, 1991), aggression (Farmer et al., 2003), stereotyping or popularity (Eder, 1985; Fishkin et al., 1993; Franzoi et al., 1994), masculinity and popularity (Pascoe, 2003)).

3.1.4. The Deviants

As opposed to the Elites, Athletes, and Academics, the Deviants cared least about schoolwork, school extracurricular activities, or future careers. They neither excelled academically nor athletically. What primarily defined the Deviants was their proneness to use tobacco, alcohol and drugs, and engage in other risky behaviors such as involvement in violence. Deviants were described as a separate group in 37 studies. Of the seven studies that did not delineate deviants, five focused on types of a priori group comparisons which would tend to group Deviants together with other groups (Demuth, 2004; Eder, 1985; Fordham & Ogbu, 1986; Miller et al., 2003, 2005), and two studies grouped Deviants in with the Elites (Dolcini & Adler, 1994; Tolone & Tieman, 1990).

3.1.5. Others

The Others category is composed of adolescent peer group names that could not be classified into any of the four other general categories. Most of these peer groups were not very distinct (e.g., 7th graders, or “Whites”) and their members did not adhere to any clearly defined set of adolescent norms. Across studies, others also were variously termed as “Normals” (e.g., Brown & Lohr, 1987; Brown, Mounts, Lamborn, & Steinberg, 1993; Clasen & Brown, 1985; Downs & Rose, 1991), “Regulars” (e.g., Fordham & Ogbu, 1986; Mosbach & Leventhal, 1988; Sussman et al., 1990; Sussman et al., 2000), “Averages” (e.g., Brown et al., 1993; Downs & Rose, 1991; Franzoi et al., 1994; Prinstein & La Greca, 2002; Poveda & Crim, 1975; Urberg, Degirmencioglu, Tolson, & Halliday-Scher, 2000), “Nobodies” (Brown & Lohr, 1987; Clark, 1962; Cohen, 1979), “Outsiders/Loners/Nerds/Outcasts” (Gotlieb, 1975), and “Floaters” (“float” from one group to another (Dolcini & Adler, 1994)). Adolescents similar to this type seem to have made up Larkin’s “Silent Majorities” (Larkin, 1979). From among the 44 studies in the dataset, only three failed to identify Others groups. These delineated only two or three general groups to describe the whole sample (Eckert, 1983; Kipke, Montgomery et al., 1997; Kipke, Unger et al., 1997).

3.2. Group names, personality, parenting, and problem-prone behavior characteristics

Table 2 provides concise summaries of the results of each paper. Some studies provided personality descriptions. The most widely studied was self-esteem. Nine studies addressed self-esteem. Differences failed to be found across groups on self-esteem in four of these studies (except that Others were found to be higher on self-esteem than the other groups in one study (Mosbach & Leventhal, 1988)), the Deviants were described as lowest in self-esteem in three studies, and the Others were described as lowest in self-esteem in two studies.

Life satisfaction was a personality-type variable that was described in five studies, favoring Elites (two studies), Athletes (one study), Academics (one study), and Others (one study). No study described Deviants as high in life satisfaction. Other personality or affective descriptors (self-reliance, self-control, and depression/anxiety) were described in three or fewer studies, and are not discussed here. In general, we could not rank groups well on the basis of the available personality findings. The most consistent set of findings across personality data portrayed the Deviants least favorably among groups.

Parenting characteristics were examined in six studies, Elites, Athletes, Academics, and Others were ascribed good (Authoritative; i.e., high parental acceptance and control) parenting in two, one, two, and one study, respectively. The Deviants were never ascribed good parenting.

Finally, problem-prone behavior characteristics were studied. A total of 14 studies investigated cigarette smoking. Of these, the Deviants were the group most likely to smoke in 13 studies, Elites were very likely to smoke in two of these studies along with Deviants, and Others were very likely to smoke in one of these studies along with Deviants. Elites were the group most likely to smoke in one study (in which a Deviant group was not identified (Dolcini & Adler, 1994)).

A total of 18 studies investigated alcohol use. Of these, the Deviants were the group most likely to drink in 15 studies, Elites were very likely to drink in four of these studies along with Deviants, Athletes were very likely to drink in one of these studies along with Deviants, and Others were very likely to drink in one of these studies along with Deviants. Elites were the group most likely to drink alcohol in two studies (Dolcini & Adler, 1994; Tolone & Tieman, 1990). In addition, one study found that Athletes were more likely to drink than Non-athletes (Miller et al., 2003). In all three of these latter studies, no separate Deviant group was delineated.

A total of nine studies investigated marijuana use. Of these, the Deviants were the group most likely to use marijuana in eight studies, Elites were very likely to use marijuana in one of these studies along with Deviants, and Others were very likely to use marijuana in one of these studies along with Deviants. Elites were the group most likely to use marijuana in the Dolcini and Adler (1994) study only.

A total of 13 studies investigated other illicit drug use. Of these, the Deviants were the group most likely to use other illicit drugs in all 13 studies, and Others were very likely to use other illicit drugs in one of these studies along with Deviants.

A total of 10 studies investigated teen dating/participation in sexual behavior. Of these, the Deviants were the group most likely to participate in teen sex in seven studies, Elites were very likely to participate in teen sex in two of these studies along with Deviants, Athletes were very likely to participate in teen sex in one of these studies along with Deviants, and Others were very likely to participate in teen sex in one of these studies along with Deviants. In three studies, in which Loners were compared to Non-loners (Demuth, 2004), Athletes were compared to non-Athletes (Miller et al., 2005), and Elites were compared with Academics and Others (Dolcini & Adler, 1994), Non-loners, Athletes, and Elites were most likely to participate in teen sex. There was no Deviant group identified in these three studies.

Finally, a total of seven studies examined involvement in violent situations (e.g., anger, fights, victimization, or stealing property). Of these, Deviants were the group most likely to participate in violent situations in six of these studies, and Others were very likely to participate in violence in one of these studies along with Deviants. In one study, in which “Socials” were compared with “Loners and Nerds” (Tolone & Tieman, 1990), Elites were more likely to participate in violence. No Deviant group was identified in that study. Other problem behaviors (i.e., drinking coffee, cruising) were reported in one study each. In both of these, Deviants were most likely to participate in the problem behavior.

4. Discussion

Moderately high agreement regarding placement of specific names into the general categories – the Elites, Athletes, Academics, Deviants, and Others – was achieved. In addition, these groups generally demonstrated the lifestyle characteristics that they depict. The Deviants perhaps were the most distinct among the groups. Self-identification as part of a Deviant group showed the greatest stability over time (Sussman et al., 1994), and self-other ratings of group identification were highest in concordance for Deviants (Urberg et al., 2000). Deviants were identified in 37 of the studies including all three studies conducted outside the U.S. Deviants were the least satisfied with life and received the worst parenting. They also were the most likely to smoke cigarettes, drink alcohol, use marijuana, use other illicit drugs, participate in dating/teen sex, and engage in violent situations.

Partial exceptions to these results are regarding alcohol use and participation in sex. On these problem behaviors, the Elites were represented as relatively high in approximately 33% of the studies. While a much lower percentage than the Deviants, who were relatively high in alcohol use and sexual behavior in at least 70% of the studies, and several of these studies involved Elites with no Deviant group identified, there was some consensus that Elites were relatively likely after the Deviants, to participate in these behaviors. These results may suggest that alcohol use and dating (sex) are associated with festive social interactions and popularity among teens and emerging adults, as well as representing a problem behavior. Conversely, cigarette smoking, marijuana use, other drug use, and violence participation may represent primarily problem behaviors.

Knowing which adolescent peer groups are most likely to engage in problem-prone behavior can help better target preventive efforts. As Gotlieb (1975) suggested, one way of looking at adolescent peer group identification is as a means by which adolescents form natural support systems without adult support or supervision. Interventions directed on peer groups as support systems are likely to be effective against propagation of unhealthy behaviors. Possibly, intervention programs focused on reducing substance use (preventing abuse) can effectively target undesirable psychosocial characteristics normative within Deviant peer groups (e.g., see the Motivation–Skills–Decision Making (MSD) model (e.g., see Sussman, Earleywine et al., 2004)). To the extent that peer group identification reflects a person perception phenomenon, one may also focus on self- and social image counteraction, to make a prosocial teen image be perceived as a more desirable type to currently self-identified Deviants. Certainly, much more research is needed to be able to more fully understand the implications of prevention programming as a function of peer group identification.

4.1. Limitations, future directions, and conclusions

There are several limitations of the peer group identification literature. Raters were not in complete agreement when placing specific groups into general categories. Some variation existed in what general types of groups one placed specific groups into, suggesting that specific groups may be viewed in different ways by different people. In addition, specific group names do change over time (e.g., the deviant group “greasers” no longer exists). Even though the same general groups exist over 40 years of studies, the variations that do exist may indicate aspects of the group identification process that cannot be learned by resorting to a finite list of general categories.

In particular, one problem is the tendency to define specific groups and then refer to an “Other” general group. Groups such as “Regulars/Averages” and “Floaters” reveal rather different if not contradictory characteristics across studies. Means to better understand youth that currently are labeled as Others is needed. In particular, this general group label really appears to us to reflect a collection of groups and future research might profitably disaggregate its constituents into meaningful and separate types.

A second more general problem is regarding how to best measure peer group identification. While participant observation is subject to observer bias, self- and peer ratings of peer group identification have their own methodological biases. As Urberg et al. [30] pointed out self-report is potentially biased because some adolescents are likely to have unrealistic perceptions about themselves. On the other hand, a peer-report method is potentially biased because peer raters may not be familiar with all of the various groups prevalent at school, their norms, and their members. Interestingly, though, Urberg et al. [30] found a high degree of association between self-identified and peer-identified groups (66%), with greater concordance for females and older adolescents. For the time being we assert that multiple measures are likely to provide a better understanding of this construct than are single measures (e.g., see Campbell & Fiske, 1959).

In addition, researchers might assess youth’s identification with multiple groups rather than the one group with which one most closely identifies. Among the 44 studies, only Kipke and colleagues permitted subjects to identify multiple groups (Kipke, Montgomery et al., 1997; Kipke, Unger et al., 1997). It is possible that youth naturally identify with many groups in different life domains (e.g., school, home, or clubs). To the extent that group identification is context dependent, one indeed could identify with multiple groups. Comparison of assessment techniques is needed for a better understanding of the parameters of group identification.

A third limitation is regarding what has really been learned about this concept over the last 40 years of research. One may argue that these groups are well-known and that all this research was not needed. Researchers simply do not yet know how this concept operates or how it might be important in youth development. Certainly, some work has indicated that one’s group identification predicts one’s lifestyle 5 years later (Sussman, Unger et al., 2004). Still, it remains unknown to what extent it operates as a social perception variable versus a group-level phenomenon, what mediates the relation between group identification and problem behaviors, or what impact on the lifespan this concept might serve. Much more progress is needed in the field than is represented by the current work (Tables 1 and ​and22).

Future research is needed to strengthen the generalizability of the construct. Except for Michell (1997) and Thurlow (2001) whose studies were completed in Great Britain, and Heaven et al. (2005), whose study was completed in Australia, all studies considered in the present review were carried out in the United States. Michell (1997) did not identify discrete peer group names as used by adolescents but classified the adolescent subjects into “Top group”, “Middle Group”, and “Bottom Group” according to the investigator’s observations. Heaven et al. (2005), and Thurlow (2001), on the other hand, found that most names used for the core groups in their samples resembled group names frequently identified in the U.S. literature, such as “Populars,” “Brains,” and “Toughs” (known as “Hard People” and “Bad People” in the British sample). Since peer group names are often used in media representing adolescent popular culture, and given its global reach, it would not be surprising to find the peer group names used in the U.S. spreading to other cultures.

Future research also should delve more into emerging adult peer group identification. Apart from Ashmore et al. (2002) who looked into adolescent peer group perceptions among undergraduates and three studies (Kipke, Montgomery et al., 1997; Kipke, Unger et al., 1997; Sussman et al., 1999) that looked into runaway/homeless youth ranging in age from 12 to 23, all the other studies were centered on adolescents. Due to this limitation present studies are likely to have failed to find out peer group names and characteristics exclusive to older youth. Emerging adults in college or elsewhere may exhibit different peer group identification patterns than younger, middle and high school students, though research is lacking.

Another future measurement direction is worth noting. Adolescents are perceived to segregate into different delineated groups, adolescents identify themselves as members of delineable groups, and teens do show a tendency to segregate into various peer groups and shape their life styles accordingly. Certainly, use of social network analysis may be important to discriminate between the perceptions youth have of groups or their own group membership versus actual group interactions, which may reflect different phenomena. For example, in a recent study of middle school youth at 16 schools in southern California, social network analysis indicated that the popular students were relatively likely to smoke (Valente, Unger, & Johnson, 2005). This result would appear contrary to group identification studies that indicate that Deviant youth, generally not considered popular, are relatively likely to smoke. Clearly, both group identification and social network data should be collected and compared to better discern how it is that group identification operates as a phenomenon. It is possible that those youth that are considered popular by Valente et al. (2005) are also those that consider themselves Deviant through a process of self-identification. We do not know this unless both types of measures are used.

In conclusion, the 44 studies included in the present review indicate that peer group identification is a phenomenon that has existed in the research literature for at least four decades. Five peer group categories emerged across studies. Although specific names of the groups within each of these categories might have changed over time, though we find little evidence of that, the characteristics have been preserved. Also, each category of peer groups is associated with a particular set of behaviors. In particular, Deviants choose to engage in health risk behaviors such as drug use and participation in violence. These kinds of choices have potential to impact the course and the quality of the adolescent’s later life. Therefore, adolescent peer group identification is a construct of consequence and demands serious attention from researchers.

Acknowledgments

This research was supported by grants from the National Institute on Drug Abuse (DA07601, DA13814, DA16094, and DA016090).

What is the function of the peer group during adolescence?

What are the four functions of the peer group?

What is a function of a peer group quizlet?

Why is peer acceptance important in adolescence?